What it does

Provides a run-code tool that executes code snippets in 40+ programming languages and returns output. Supports JavaScript, Python, PHP, Go, Ruby, Perl, C#, TypeScript, Scala, Swift, Kotlin, Julia, Haskell, Elixir, Clojure, Racket, and others. Execution requires interpreters or compilers for target languages to be installed and available in the system PATH.

Who it's for

Developers and AI agents embedding code execution into interactive workflows. Useful for teams testing language features, validating logic, or building AI systems that run code as part of reasoning. Not suitable for running untrusted code — sandboxing is the caller's responsibility.

Common use cases

• Execute JavaScript or Python snippets to test code logic without leaving your editor
• Run shell commands to query system information
• Validate code behavior across multiple programming languages
• Allow Claude to execute and verify code interactively

Setup pitfalls

• Requires language interpreters and compilers in PATH — verify with which python or node --version before use
• On Windows, npx may fail to connect; use bunx or wrap with cmd /c as a workaround
• Last commit 104 days ago; check GitHub for compatibility with your target language versions